A discrete stochastic model for performance prediction of roofing systems

A stochastic modeling of the performance of roofing components and systems is proposed using a discrete Markov chain that predicts the incremental deterioration of roofing components. The capability to predict the performance of roofing-components and systems throughout their service lives is essential and constitutes the first step towards the development of a comprehensive building envelope life-cycle asset management system. The proposed Markovian model allows the simultaneous consideration of the time-dependence and randomness of the performance of roofing components and systems taking into account the uncertainty and variability associated with material properties, environmental degradation factors, quality of workmanship, and maintenance conditions. Depending on the controlling explanatory variables, transition probability matrices are generated for roofing components and systems using in-field performance data obtained from inspections. This probabilistic model can then be coupled to an optimization algorithm for the development of a risk-based optimal maintenance management system for roofs at both project and network levels.